 |
|
Date: December 26, 1995
From: Medical Officer,
Epidemiology Branch,
Division of Market Studies (HFS-728)
Subject: Review of Study Entitled
"Measurement of Selected Fecal Parameters
in Subjects Consuming Increasing Levels
Of Olestra" (FAP 3997, Volume 280)
To: Helen Thorsheim, Ph.D.,
Consumer Safety Officer,
Office of Premarket Approval, (HFS-216)
Through: Raymond E. Schucker, Ph.D., Director,
Division of Market Studies
Study Objectives: The purposes of this study were to quantitate the frequency, duration, and intensity of pre- defined GI symptoms as a function of dietary level of olestra, and to quantitate selected fecal parameters that might change in response to changes in olestra intake.
Study Design: The study consisted of two phases. A screening phase was conducted to identify subjects who reported GI symptoms in response to olestra consumption. This was followed by the study phase during which the identified subjects were given increasing dietary levels of olestra and GI symptoms were recorded and fecal measurements were made.
Screening Phase
This was a four-week, cross-over design with two treatment groups, 0 and 20 g/d olestra. The olestra substituted for 20 g of triglyceride in the three daily meals. Roughly one- third of the daily dose was provided in each meal.
Fifty-two adult subjects who had reported GI symptoms in previous product acceptance or olestra clinical studies were acclimated to the study procedures during a three-day baseline period in which they were given placebo meals. They were then divided into two groups and given placebo meals or meals which provided 20 g/d olestra for five days. After a seven-day washout period, the subjects were again given placebo meals (containing triglycerides) for three days, and then crossed- over to olestra or placebo meals for five days. After the second treatment period, the subjects were monitored for a four-day washout period.
The subjects were fed one of four meal sizes, based on four average body weights: 50, 65, 80, and 95 kg. The meals provided 30 kcal of energy per kg of body weight. The olestra was incorporated into the meals in french fries, ice cream,
Page 2 - Helen Thorsheim, Ph.D.
margarine, and cookies. The subjects were fed all meals during the baseline, treatment, and washout periods under supervision at the clinical site. The stiffness of the olestra was approximately 90 kPa/s.
The frequency, duration, and severity of nine pre-defined GI symptoms were documented daily by the subjects, starting at the beginning of the baseline period and continuing through the final four-day washout period. Diarrhea was defined as "excessive frequency of very loose/watery stools that are extremely difficult or impossible to control." Loose stools were defined as "a bowel movement that is easier to pass than normal, but is not watery and unformed."
At the completion of the screening phase those subjects who reported an increase in the frequency, severity, or duration of GI symptoms during the olestra period, relative to the placebo period, were selected to take part in the study phase. Eighteen subjects met the selection criteria.
Study Phase
This phase was a crossover, placebo-controlled, single-blind (subject) design with three treatment groups, 0, 10, and 20 g/d olestra. Each subject received each treatment for seven days. The treatment periods were separated by seven-day washout periods.
During the treatment periods, the subjects were fed all meals under supervision at the clinical site. The subjects were fed according to their weight, 30, kcal/kg of body weight/day. Portions of food were weighed before the meals and uneaten amounts were weighed after the meal to monitor consumption. The olestra dose was divided among the three daily meals and was incorporated into the meals in french fries, ice cream, margarine, and cookies. The placebo and olestra-containing meals provided the same amount of metabolizable energy. During the treatment periods, the subjects were not allowed to eat any food other than that provided in the test meals. During the washout periods, subjects ate their habitual diets at home.
GI symptoms were ascertained during the treatment periods and the first four days of the washout periods by GI Assessment Records completed daily by the subjects. For each GI symptom episode, the subject recorded the date, the time of day, and the intensity. The intensity scale for GI symptoms was graded as follows: 0 (none); 1 (slight); 2 (mild); 3 (moderate); and
Page 3 - Helen Thorsheim, Ph.D.
4 (severe). Total fecal collections were made for the last three days of each treatment period and the daily collections were pooled. To complete the study and have data included in the analyses, a subject had to provide at least one fecal sample for each three-day collection period.
Subject selection
Normal healthy males and females (18 to 60 years of age) were selected for the study from a population of subjects who had reported GI symptoms while consuming olestra in previous product acceptance studies. In addition, to be selected for the study phase, a subject had to report an increased frequency, severity, or duration of GI symptoms while on the 20 g/d olestra treatment, relative to the placebo treatment.
Stools were collected into plastic containers and immediately frozen at -20 degrees C. Sealable plastic bags and an ice chest containing dry ice were provided to the subjects for collection of stools at home. These samples were transferred to -20 degrees C as rapidly as possible. All fecal samples were kept frozen until analyzed.
Fecal samples were initially collected on an individual basis for each bowel movement and several measurements were made (e.g., wt weight, volume, density). Subsequently, all fecal samples from the three-day collection period were pooled (thus diarrheal, loose, and normal stools were pooled) for analyses of the remaining parameters (e.g., water concentration, dry weight, olestra analyses, Na, K, Cl, total and individual bile acids, free fatty acids, triglycerides, and total lipids by the Mayo Medical Laboratories (Rochester, MN). Because the analyses of stool electrolytes were conducted on pooled samples from the three-day collection period, separate daily values for an individual are not available.
Results
More than 99% of the olestra-containing test foods were consumed during both the 10 g/d olestra and 20 g/d olestra treatments.
Results: Screening period
Of the 52 subjects who enrolled in the study, 47 completed the screening phase. None of the five subjects who withdrew did so because of medical reasons. Ten different subjects reported adverse experiences during the screening phase, most commonly GI in nature. Two subjects reported diarrhea during the placebo period; one of these two reported loose stools also. During the olestra consumption period, one subject reported diarrhea and another reported loose stools.
Page 4 - Helen Thorsheim, Ph.D.
Results: Study phase: Incidence of diarrhea and loose stools
Of the 47 subjects who completed the screening phase, 18 were selected to participate in the study phase. Two subjects withdrew prior to start of treatment, and one subject withdrew after the first treatment period (20 g/d olestra); none of the withdrawals was due to medical reasons. Therefore, 15 subjects completed the placebo arm of the study phase and the olestra 10 g/d arm, and 16 subjects completed the olestra 20 g/d arm of the study.
 |
Figure 1 (right) summarizes the number of study subjects who reported diarrhea or loose stools, by olestra dose. There was a steady increase in the number of subjects who reported diarrhea with increasing dose of olestra consumed. Three (20%) subjects reported diarrhea while eating 0 g/d olestra; 6 (40% while eating 10 g/d olestra; and 11 (69%) while eating 20 g/d olestra. The difference in incidence of reported diarrhea during placebo and 20 g/d olestra treatment was statistically significant (p=0.004, Mantel-Haenzel chi square test). There was also an increase in the number of subjects reporting loose stools with increasing olestra dose, but this increase did not reach statistical significance (p=0.08 comparing placebo vs. 20 g/d olestra).
Figure 2 summarizes the number of diarrheal or loose stool episodes, by olestra dose consumed. There was a steady increase in the number of diarrheal bowel movements with increasing olestra dose eaten: 5 episodes of diarrhea among subjects while eating 0 g/d olestra; 14 episodes of diarrhea among subjects whole eating 10 g/d olestra; and 40 episodes of diarrhea among subjects while eating 20 g/d olestra. The mean number of diarrheal bowel movements per subject reporting diarrhea during the study phase increased steadily with increasing olestra dose consumed; 1.7 diarrheal bowel movements per subject reporting diarrhea during placebo treatment [5 episodes of diarrhea among 3 subjects]; 2.3 diarrheal bowel movements per subject reporting diarrhea during 10 g/d olestra treatment [14 episodes of diarrhea among 6 subjects]; and 3.6 diarrheal bowel movements per subject reporting diarrhea during 20 g/d olestra treatment [40 episodes of diarrhea among 11 subjects]. The number of subjects experiencing diarrhea as reportedly "severe" increased with increasing doses of olestra consumed: none of the three subjects who experienced diarrhea while eating placebo reported the diarrhea as "severe"; two of the six subjects who experienced diarrhea while eating 10 g/d olestra reported the diarrhea as "severe"; and six of the 11 subjects who experienced diarrhea while eating 20 g/d olestra reported the diarrhea as "severe". The number of episodes of loose stools also increased with increasing amounts of olestra consumed.
Page 5 - Helen Thorsheim, Ph.D.
In summary, both the incidence of diarrhea and the mean number of diarrheal bowel movements per subject reporting any diarrhea increased with increasing olestra consumed. Additionally, there was a steady increase in the incidence of diarrhea reported as "severe" with increasing olestra consumed.
Results: Fecal measurements during study phase
Six subjects reported a total of 13 episodes of diarrhea during the stool collection periods (i.e., the last three days of each treatment period). All episodes of diarrhea that were reported during the stool collection periods occurred among subjects eating either olestra 10 g/d or olestra 20 g/d. Summarized below are all reported bowel movements reported for the six subjects who experienced diarrhea during the stool collection periods; bowel movements not collected are noted.
Placebo Stool Collection Period
Subject #103 1 episode of loose stools on day #5
1 episode of loose stools on day #6
1 episode of loose stools on day #7
Subject #125 1 episode of normal stools on day #5
1 episode of loose stools on day #6
1 episode of normal stools on day #6
1 episode of normal stools on day #7
Subject #0139 1 episode of normal stools on day #5
1 episode of loose stools on day #6
1 episode of normal stools on day #7
Subject #0146 1 episode of normal stools on day #5
1 episode of normal stools on day #7
Subject #0152 2 episodes of loose stools on day #5
1 episode of loose stools on day #6
1 episode of loose stools on day #7
Olestra 10 g/d Stool Collection Period
Subject #0103 1 episode of loose stools on day #5
2 episode of diarrhea on day #7
(one of these specimens was not collected)
Subject #0125 1 episode of normal stools on day #5
1 episode of loose stools on day #6
2 episode of diarrhea on day #7
Page 6 - Helen Thorsheim, Ph.D.
Subject #0139 1 episode of normal stools on day #5
1 episode of normal stools on day #6
Subject #0146 1 episode of normal stools on day #5
2 episodes of normal stools on day #6
1 episode of normal stools on day #7
Subject #0152 1 episode of loose stools on day #5
1 episode of normal stools on day #6
1 episode of loose stools on day #7
(this specimen was not collected)
Olestra 20 g/d Stool Collection Period
Subject #0103 2 episodes of loose stools on day #5
1 episode of diarrhea on day #6
1 episode of loose stools on day #7
Subject #0122 3 episodes of diarrhea on day #5
2 episodes of diarrhea on day #6
1 episode of loose stools on day #7
Subject #0125 1 episode of normal stools on day #5
1 episode of normal stools on day #6
2 episodes of loose stools on day #6
1 episode of loose stools on day #7
(this specimen was not collected)
Subject #0139 1 episode of diarrhea on day #5
1 episode of loose stools on day #6
1 episode of normal stools on day #7
Subject #0146 1 episode of diarrhea on day #5
1 episode of loose stools on day #6
2 episodes of normal stools on day #6
Subject # 0152 2 episodes of loose stools on day #5
1 episode of diarrhea on day #6
1 episode of loose stools on day #6
(this specimen was not collected)
2 episodes of loose stools on day #7
(one of these specimens was not collected)
Thus, among the six subjects who reported diarrhea during the fecal collection period, there were five instances of stool samples not being collected. The ramifications of these omitted stools is discussed below. Summarized below are the wet weights of stool (includes the 10 or 20 g/d of nonabsorbed olestra) for subjects who reported diarrhea and for subjects who did not report diarrhea during
Page 7 - Helen Thorsheim, Ph. D.
the three days stools were collected; this data is abstracted from Exhibit 5.b. The mean wet weight of stool for persons who consumed olestra and reported diarrhea during the fecal collection periods ([140± 36 for olestra 10 g/d] and [158±36 for olestra 20 g/d]) exceeded the mean wet weight of their stools during the placebo period (111±61), when none of the subjects reported diarrhea.
Mean (± SD) wet weight (g/24 hr) of stools, by olestra dose, for subjects who reported diarrhea and for subjects who did not report diarrhea during the three days stools were collected.
Placebo 10 g/d Olestra 20 g/d Olestra
Non-diarrhea Non-diarrhea Diarrhea Non-diarrhea Diarrhea
stool stool stool stool stool
Subjects (n=6) reporting diarrhea during fecal collection periods
111±61 88±39 140±36 None 158±36
(n=3) (n=2) (n=6)
Subjects (n=10) reporting diarrhea during fecal collection periods
89±67 79±36 None 79±76 None
Regarding subjects who reported diarrhea during the stool collection period, I asked Curtis Barron, Ph.D. (Division of Mathematics) to conduct a paired T-test to determine whether the wet weight of stools from subject who reported having diarrhea while consuming 20 g/d olestra differed from the weight of nondiarrheal stools during the placebo period. Data for this t-test calculation was obtained from Exhibit 5.b. (This analysis was not done for the 10 g/d olestra subjects since only two subjects reported diarrhea on this dose of olestra.) Subject #0122 was excluded from the analysis since he withdrew from the study without going through the placebo treatment period; thus, analysis was restricted to the paired results from five subjects. The mean daily weight of stools while consuming 20 g/d olestra was greater than that of stools while on placebo (T statistic 2.715, p-value 0.053). This value underestimates the true difference since three stool
Page 8 - Helen Thorsheim, Ph.D.
episodes reported by two subjects during the olestra 20 g/d group (see above) were not collected. Results of the paired T-test after subtracting the weight of olestra (20g) from the diarrheal stool weights for each subject yielded a p-value of 0.17.
In my opinion, the results of the paired T-test suggest that olestra-associated diarrheal stools were heavier than non- diarrheal stools (p=0.053). The presence of nonabsorbed olestra in the stool accounts for some, but not all, of this increase in stool weight. I assume that the remainder of this difference in weight is due to increased water in the diarrheal stool of olestra consumers.
The following values are the maximum stool weights for the five subjects who reported having diarrhea during the stool collection period and for whom data exist for both the placebo and the olestra 20 g/d treatment period:
Subject # Maximum daily stool Maximum daily stool g/d weight on placebo weight on olestra 20 103 (female) 213 279 125 (female) 281 456 139 (male) 70 170 146 (female) 115 257 152 (female) 103 281
This data suggests that, at least on some days while consuming olestra 20 g/d, stool weights for all four females exceeded the cutoff weight for diarrheal stool (235 g/d for men; 175 g/d for women) (reference provided by petitioner: Fine et. al.).
The stool weight data from the six subjects who reported diarrhea during the collection periods also suggests that there is a physiologic difference in the stools that subjects report as "diarrhea" versus that they call "loose" versus that they call "normal". The evidence for this is that the mean daily weight of stools described as "diarrhea" by persons consuming 20 g/d olestra was 177 g; the mean daily weight of stools described as "loose" for subjects consuming 20 g/d olestra was 137 g; by comparison, the mean daily weight of stools described as "normal" by the five subjects was 80 g. This data is summarized in the table below.
Page 9 - Helen Thorsheim, Ph.D.
This data is summarized in the table below.
Mean daily weight of stools, by subjective description of stools, for subjects who reported having diarrhea during stool collection period
Stool Type Mean daily weight (grams) "Diarrhea" 177 "Loose stools" 137 "Normal stools" 80
Data presented in Appendix G regarding the stool water concentration -- expressed as a percent of stools by weight -- suggests that the stool water concentration of subjects who reported having diarrhea during the olestra 20 g/d period did not differ from that of their nondiarrheal stools during the placebo period. However, even though the percent of stools composed of water may not have differed, it is possible that water loss was greater in subjects with olestra-associated diarrhea because of the greater mass (weight) of stool passed. The data provided regarding stool electrolytes is presented in meq/day (Exhibits 7, 8, and 9), and suggests that the electrolyte concentration of persons eating olestra and reporting diarrheal stools did not differ from that of their nondiarrheal stools during the placebo period. As noted above, it was not possible to analyze stool electrolyte values for subjects by individual stools or by individual days since the stools were pooled from the three-day collection period prior to analysis for electrolytes.
I am not able to analyze the section on stool lipids without further information. Specifically:
1. Do the values in Appendix G under the column entitles "feces total lipid" include the weight of olestra?
A. If not, why does total feces lipid increase with olestra ingestion? In other words, does olestra interfere with lipid absorption in the gut?
Page 10 - Helen Thorsheim, Ph.D.
B. If feces total lipid does include olestra, why are some values well below the putative olestra dose? Does this imply non-compliance with diet (e.g., subject #102, 117, and 119 on olestra 10 g/d diet; and subject #102, 117, 130, 131, 147, and 150 olestra 20 g/d diet)? I note that none of these individuals experienced diarrhea during the respective treatment periods.
Additional request: Please provide the results stool analyses for olestra.
Conclusions
Recognizing that this study was small and was based on a total of six individuals who experienced diarrhea during the stool collection periods, listed below are my conclusions:
1. As demonstrated by the paired t-test, in aggregate, subjects who consume olestra 20 g/d experience mean daily stool weights that exceed the weight of their stools while eating placebo.
2. Two lines of evidence suggest that the increased mean daily stool weights among persons reporting diarrheal stools while consuming 20 g of olestra per day meets the definition of diarrhea according to Fine et. al. (reference provided by petitioner). First, at least on some days while consuming olestra 20 g/d, stool weight for some subjects exceeded the cutoff weight for diarrheal stool (175 g/d for women). Second, the observed mean daily stool weights by type of stool subjectively experienced suggest that there is a physiologic difference in the stools that subjects report as "diarrhea" versus that they call "loose" versus that they call "normal".
3. Data presented in Appendix G regarding the stool water concentration -- expressed as a percent of stools by weight -- suggests that the stool water concentration of subjects who reported having diarrhea during the olestra 20 g/d period did not differ from that of their nondiarrheal stools during the placebo period. However, even though the percent of stools composed of water may not have differed, it is possible that water loss was greater in subjects reporting olestra- associated diarrhea because of the greater mass (weight) of stool passed.
4. It is important to note the high incidence of diarrheal episodes reported by those six subjects who experienced this symptom during the 20 g/d olestra period: 40 episodes of diarrhea among 11 subjects. This is particularly noteworthy in light of the definition of diarrhea provided to the subjects: "excessive frequency of very loose/watery stools that are extremely difficult or impossible to control."
Page 11
Helen Thorsheim, Ph.D. My conclusion is that the increased water loss in the stools of subjects reporting olestra-associated diarrhea or loose stools is of concern. The concern is not so much for young, healthy persons, but for the elderly and young children. The elderly are more likely than young, healthy persons to have underlying medical conditions that could be exacerbated if they become dehydrated (e.g., orthostatic hypotension leading to falls in an elderly person with cardiovascular disease). In addition, the elderly are more likely than young, healthy persons to be taking other medications that can have diarrhea as a side effect. For example, the Physicians Desk Reference- Drug Interactions and Side Effects Index lists 654 medications that can cause diarrhea (with frequencies raging from less than 1 percent to 83 percent).
These conclusions reinforce the need to have a clear, informative label on all olestra-containing products. They also strongly suggest that aggressive post-marketing surveillance is required should olestra be approved so that the incidence of potential adverse effects can by monitored in subpopulations that were not studied at all prior to approval (i.e., the elderly) or in whom studies were very brief (i.e., children).
 |
source: http://www.cspinet.org/olestra/fda_memo.html 24oct01
|
If you have come to this page from an outside location click here to get back to mindfully.org |